Dictation, file transcription or a local model
Transcription tools do three different jobs. Pick the kind first, then the tool:
| Kind | Examples | Good for |
|---|---|---|
| Live dictation | Google Docs voice typing, Windows voice typing, Dictation on Mac, Speechnotes | Speaking text into a document instead of typing |
| Online file transcription | Riverside, Otter.ai, Notta, Descript | Recordings, interviews, meetings and podcasts |
| Local models | Whisper, whisper.cpp, Buzz, Vosk | Transcribing on your own computer without a minute plan |
| Manual helper | oTranscribe | Typing a transcript yourself with easy playback controls |
Free minutes compared
Online services count minutes in different ways: per month, per conversation or per uploaded file. These are the free limits we found on the official pages:
| Service | Free limit |
|---|---|
| Otter.ai Basic | 300 minutes a month, up to 30 minutes per conversation, 3 file imports over the life of the account |
| Notta Free | 120 transcription minutes a month, up to 3 minutes per conversation, 50 file uploads a month |
| Descript Free | 60 media minutes a month; video exports carry a watermark |
| Riverside | Unlimited transcriptions, no sign up, per its transcription page |
| Speechnotes | Unlimited dictation with ads; file transcription costs US$0.10 a minute |
Languages differ too. Otter.ai lists English, Spanish, French, German, Japanese and Chinese, Descript lists 25 languages and Riverside says it covers 100+ languages.
Transcribe on your own computer
If you have many hours of audio or private recordings, an open source model on your own machine avoids monthly minute caps. Whisper by OpenAI is released with code and model weights under the MIT License. It runs with Python and ffmpeg, and its README lists about 1 GB of video memory for the smallest models and about 10 GB for the large one.
You do not need to write code. Buzz is a desktop app that transcribes and translates audio offline with Whisper, and whisper.cpp runs the same model in C and C++ on many systems. Vosk is an offline toolkit for 20+ languages that also runs on phones and a Raspberry Pi.
If you prefer to type the transcript yourself, oTranscribe gives you playback controls and timestamps in the browser and says your audio file and transcript never leave your computer.
Prepare the audio first
A clean file saves minutes and gives a better transcript:
- Cut long silences and off-topic parts before you upload, so they do not use up free minutes. Our trim audio tool cuts the start and end of a file.
- If you only have a video, save the sound track with extract audio from video.
- Check the file type. Riverside lists MP3, WAV, MP4 and MOV. Our audio converter changes the format in your browser.
- Read the result once against the recording. Names, numbers and technical terms are where automatic transcripts most often need a fix.
Free audio tools on this site
- Speech to text: dictate with your microphone and copy the text.
- Voice recorder: record audio and download it, nothing uploaded.
- Extract audio from video: save the sound of a video as an audio file.
- Trim audio: cut the start and end of a recording with a waveform.
- Audio converter: convert between MP3, WAV, M4A, OGG and more.
For recording your screen and voice, see our list of free screen recorders. We opened each tool's official page on the date in the table and wrote down only what it says. "Not stated" means the page does not say it.
Frequently asked questions
Which free transcription service gives the most minutes?
Otter.ai Basic gives 300 minutes a month, but each conversation is capped at 30 minutes and you can import only 3 files in total. Notta Free gives 120 minutes a month and Descript Free 60 media minutes. Riverside says its transcription tool is unlimited and needs no sign up.
Can I transcribe audio offline for free?
Yes. Whisper, whisper.cpp, Buzz and Vosk are open source and run on your own computer. Buzz is the easiest start because it is a desktop app for macOS, Windows and Linux; on a Mac it needs Apple silicon.
How do I transcribe a video file?
Riverside accepts MP4 and MOV files as well as MP3 and WAV. For other services, save the sound track first with the extract audio from video tool on this site and upload the audio file.
Is voice typing on Windows and Mac free to use?
Both are built into the system. On Windows, press the Windows key and H; Microsoft says voice typing needs an internet connection. On a Mac, Apple says you can dictate text of any length, and Dictation stops after 30 seconds without speech.
Which languages do free transcription tools support?
It depends on the tool. Otter.ai lists six languages, Descript 25, Riverside 100+ and Vosk 20+ languages and dialects. Google Docs voice typing supports many languages, but its voice commands work only in English.